Why Gmail Spam Filters Miss So Much and What to Add

Google claims Gmail blocks 99.9% of spam. For a service processing hundreds of billions of emails per year, that is genuinely impressive. But 0.1% of a very large number is still a large number. If you receive 200 emails per day, 99.9% accuracy means one spam message every five days. For a business address, the volume is higher and the false negatives are more frequent.

The bigger problem is not the percentage — it is which messages get through. Gmail’s spam filter is not randomly missing 0.1% of spam. It consistently misses certain types of messages because of structural decisions in how it works. Understanding those blind spots tells you exactly what to add on top.

What Gmail catches well

Credit where it is due. Gmail is genuinely excellent at catching:

• Mass spam campaigns: Identical or near-identical messages sent to millions of users. Gmail’s scale means it sees these campaigns early and blocks them fast.
• Known bad senders: IP addresses and domains with poor reputation get blocked before the message is even processed.
• Phishing templates: Known phishing patterns — fake login pages, package delivery scams, account verification traps — are flagged reliably because Gmail has seen millions of examples.
• Obvious content patterns: ALL CAPS subjects, suspicious attachments, keyword-stuffed messages, and other traditional spam signals.

For the average personal user with one Gmail address, this is often enough. The problems start when your email usage gets more complex. For a full breakdown of how the detection works, see How Gmail Detects Spam.

Where Gmail consistently fails

Forwarded email

This is Gmail’s biggest blind spot for custom domain users. When you forward email from another service to Gmail, the spam filter has less to work with. The sending IP is the forwarding server (which is legitimate), SPF checks fail or become irrelevant, and the forwarding service may strip or modify headers that Gmail uses for scoring.

The result: spam that Gmail would catch if sent directly gets forwarded past its filters. If you use a custom domain with email forwarding to Gmail, you have probably noticed this. The forwarding service sends everything — spam included — and Gmail treats it as semi-trusted because the forwarding server has a clean reputation. For more on why this happens, see Why Email Forwarding Breaks SPF.

Newsletter and promotional gray area

Gmail’s Promotions tab was designed to solve this, but it created a new problem. Legitimate newsletters you signed up for land in Promotions, where you might never see them. Meanwhile, borderline marketing spam — messages from companies you never opted into — often lands in your primary inbox because the sender has good authentication and a clean domain reputation.

Gmail does not distinguish between “marketing email you wanted” and “marketing email you did not want.” It classifies by format and sender type, not by whether you actually subscribed. The result is that wanted newsletters get buried and unwanted marketing gets through.

Compromised account spam

When a legitimate email account is compromised and used to send spam, Gmail’s reputation systems work against it. The sending domain has years of legitimate history. The IP belongs to a trusted email provider. SPF, DKIM, and DMARC all pass. The spam comes from a “clean” source, and Gmail’s infrastructure-level checks see nothing wrong.

Content analysis should catch these, but compromised account spam is often more sophisticated — the attacker has access to the account’s sending history and can mimic the writing style or reply to existing threads.

AI-generated messages

Traditional spam had telltale signs: poor grammar, obvious templates, suspicious formatting. AI-generated spam has none of these. Each message is unique, grammatically perfect, and contextually relevant. Gmail’s content filters, trained on patterns from older spam, struggle with messages that look exactly like legitimate email.

Google is adapting, but the arms race favors the attacker: generating unique spam is cheap, while training detection models on new patterns takes time and data.

Targeted spear phishing

Gmail’s strength is scale — it catches spam that millions of users receive. Spear phishing targets individuals or small groups with customized messages. There is no volume pattern to detect, no template to match, and no mass-complaint signal to trigger blocking. Gmail relies on content heuristics for these, which produce more false negatives than mass-spam detection.

Subscription bombing

An attacker signs your email address up for hundreds of legitimate mailing lists simultaneously. Each individual email is from a real company with good authentication — Gmail has no reason to block them. The volume overwhelms your inbox, potentially hiding a password reset or account takeover notification in the flood. Gmail has no mechanism to detect coordinated sign-up attacks because each message is individually legitimate. See Subscription Bombardment Explained for more on this attack.

What “Report Spam” actually does (and does not do)

When you hit “Report Spam” in Gmail, you are contributing to a collective signal. If enough users report the same sender or pattern, Gmail adjusts its filters. This works well for mass campaigns.

But it does not help with:

• Targeted spam: If only you receive a message, your single report has minimal impact on the global model.
• New senders: The report helps future filtering, but the current message already got through.
• Forwarded spam: Reporting trains Gmail against the forwarding server, not the actual spammer. You may end up causing legitimate forwarded email to be flagged.

For more details on what the report button actually triggers, see What Happens When You Mark Email as Spam.

What to add on top of Gmail

Pre-filter before forwarding

If you forward email from a custom domain to Gmail, the single biggest improvement is filtering before the forward happens. An MX relay or filtering service that sits between the internet and your forwarding setup catches spam before Gmail ever sees it. Gmail’s blind spot for forwarded email disappears because spam never gets forwarded in the first place.

Email aliases for compartmentalization

Instead of giving out your real Gmail address, use email aliases. Each service gets a unique address. When one starts receiving spam, you disable it without affecting the rest of your email. This does not filter spam — it prevents it from reaching your real address at all. See Find Out Who Sold Your Email for how aliases help identify the source of spam.

Sender authentication for your custom domain

If you own a custom domain, properly configured SPF, DKIM, and DMARC do not just protect others from spoofed email claiming to be from you — they also improve how your domain’s email is handled by receiving servers. A domain with a p=reject DMARC policy tells the world that unauthenticated email from your domain is fake. For setup details, see SPF, DKIM, and DMARC Explained.

DNSBL-level blocking

Gmail uses its own reputation data, but it does not publicly disclose which DNSBL lists it checks. Adding a layer that explicitly checks Spamhaus, Barracuda, and SpamCop gives you known-good reputation data that catches senders Gmail might not flag. See DNSBL Providers Compared.

AI-based content analysis

A second opinion on content analysis catches what Gmail misses. Different AI models, trained on different data with different approaches, have different blind spots. Running email through an independent classifier in addition to Gmail’s built-in filtering reduces the chance that a message passes both systems.

How Cleanbox fills Gmail’s gaps

Cleanbox is designed to work alongside your existing email, including Gmail. There are two approaches:

As a forwarding layer: Point your custom domain’s MX records at Cleanbox. Every incoming email passes through Rspamd scoring, DNSBL checks, SPF/DKIM/DMARC verification, ClamAV scanning, and AI classification before being forwarded to Gmail. Spam is caught before forwarding, so Gmail’s forwarded-email blind spot never comes into play.

As an MX relay: For domains using Cleanbox Relay, email is filtered and delivered directly to your mail server. Gmail is not in the path at all — you get the full filtering stack without depending on Gmail’s spam detection.

Both approaches give you what Gmail does not: per-address spam thresholds, quarantine review for borderline messages, detailed spam reports explaining why each message was flagged, and multiple independent detection layers that catch different types of spam. For a broader comparison of filtering options, see Best Spam Filters for Gmail and Outlook and our Gmail vs Outlook comparison.